Assessing GPT-4’s diagnostic accuracy with darker skin tones: underperformance and implications
Clinical and Experimental Dermatology · 18.09.2024 · 49 (10) · 1244-1245
DOI: 10.1093/ced/llae158 ↗Kurz gefasst
Leserbrief in der Zeitschrift Clinical and Experimental Dermatology, online erschienen am 02.05.2024. Verfasst von Edgar Akuffo-Addo und 6 weiteren Autor:innen. Beteiligte Einrichtung: Division of Dermatology, Department of Medicine, University of Toronto, Toronto, ON, Canada. Hauptthemen laut MEDLINE: Artificial Intelligence, Skin Pigmentation. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 5 Literaturangaben.
Abstract (Originalsprache)
This study aimed to assess the accuracy of GPT-4 in generating appropriate differential diagnoses and arriving at the correct diagnoses for common skin lesions. Additionally, we investigated any differences in its diagnostic accuracy between darker and lighter skin tones. GPT-4 exhibited better performance in providing the correct diagnosis for lighter skin tones (44%, n = 11 of 25) compared with darker skin tones (12%, n = 3 of 25), and this was statistically significant (P < 0.05). Furthermore, with each unit increase in the Fitzpatrick scale, GPT-4’s performance decreased by 11.4% in accurately providing a differential diagnosis and by 7.1% in accurately providing the correct diagnosis.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Division of Dermatology, Department of Medicine, University of Toronto, Toronto, ON, Canada.
- Rowan School of Osteopathic Medicine, Stratford, NJ, USA.
- University of Texas Medical Branch, Galveston, TX, USA.
- Medical College of Georgia, AU/UGA Medical Partnership, Athens, GA, USA.
- Department of Medicine, Division of Dermatology, David Geffen School of Medicine, University of California, Los Angeles, CA, USA.
- University of Nevada, Reno School of Medicine, Reno, NV, USA.
ORCID-Profile: Edgar Akuffo-Addo
Zitieren
APA
Akuffo-Addo, E., Samman, L., Munawar, L., Akbik, M., Kokikian, N., Wescott, R., & Wu, J. J. (2024). Assessing GPT-4’s diagnostic accuracy with darker skin tones: underperformance and implications. Clinical and Experimental Dermatology, 49(10), 1244-1245. https://doi.org/10.1093/ced/llae158
Vancouver
Akuffo-Addo E, Samman L, Munawar L, Akbik M, Kokikian N, Wescott R, et al. Assessing GPT-4’s diagnostic accuracy with darker skin tones: underperformance and implications. Clinical and Experimental Dermatology. 2024;49(10):1244-1245. doi:10.1093/ced/llae158
BibTeX
@article{AkuffoAddo2024,
author = {Akuffo-Addo, Edgar and Samman, Luna and Munawar, Leena and Akbik, Maya and Kokikian, Nelly and Wescott, Raquel and Wu, Jashin J},
title = {Assessing GPT-4’s diagnostic accuracy with darker skin tones: underperformance and implications},
journal = {Clinical and Experimental Dermatology},
year = {2024},
volume = {49},
number = {10},
pages = {1244--1245},
doi = {10.1093/ced/llae158}
}
EUVIMED